Types of Machine Learning

In the previous post, we looked at what is meant by Machine Learning and now we will take a look at the types of Machine Learning.

Machine Learning is basically of 3 types :

1. Supervised Machine Learning
2. Unsupervised Machine Learning
3. Reinforcement Machine Learning

Supervised Machine Learning:

The meaning of the word Supervised is any task or activity which is done under the observation of someone. 

In Supervised machine learning the same thing happens, we provide the machine learning model with some example data and train the model based on these examples. 

When training is completed, our model can easily predict the new data which is never seen by the model.

For example: When detecting whether an email is spam or not, we can provide the model sample data which tells the model the type of emails and whether they are spam or not. So when a new email is introduced to the model for classification, it looks at the previous learned examples and mark the email as spam or not.

The most widely used Supervised learning algorithms are:

• Linear Regression
• Support Vector Machines
• Decision Trees
• Logistic regression
• Neural Networks   

Unsupervised Machine Learning:

The meaning of the word unsupervised is not done or acting under supervision.

In Unsupervised Machine learning, the model is provided data that is not labeled, and the model itself lookout for patterns in the data and recognizes those patterns. 

The model based on these patterns make groups or clusters or boundaries along the data. Whenever any new data is provided to the model, it look out for the already defined patterns and put the data in clusters or groups based on these patterns.

For example, Suppose you are the owner of a grocery store and you want to analyse the customers that visit your store, so you upload the list of visitors along with some other data to the model. What model will do, it will look for some relationship and then group the visitors into different groups or categories.

Some of the types of Unsupervised Learning algorithms are 

• K-means Clustering
• Principal Component Analysis

Reinforcement Machine Learning:

The meaning of the word Reinforcement is the process of encouraging or establishing a belief or pattern of behavior.

In Reinforcement Learning, the agent is encouraged to form a pattern or take decisions based on the previously taken decisions. 

The agent must try various actions and based on these actions, the agent is rewarded or punished.

The ultimate goal of the agent is to maximize the cumulative reward. Here the end goal is defined and agent must achieve that goal by taking the decisions to reach that goal.

For example: While training an agent in a game to compete with humans, the end goal of the agent is to defeat the human player and be a winner. So to achieve the same, the agent can take decisions and decide for itself whether it is moving toward the goal or away from it.

Q-learning is a widely used Reinforcement learning algorithm